OpenAI Launches GPT-6.1 Sol for Programming and Complex Tasks

Washington, October 1, 2026 : OpenAI has launched GPT-6.1 Sol, a new AI model designed to handle programming and other complex tasks with greater accuracy and advanced reasoning capabilities.

In a statement, the company said the model recorded a lower error rate in real-world accuracy tests using low-level reasoning, falling to 7.7 percent from 11.4 percent for the previous GPT-6 Sol model. OpenAI said its performance is approaching that of GPT-6 Astra.

GPT-6.1 Sol is designed to support AI agents in carrying out programming tasks beyond simply generating code suggestions. The model can assist with planning tasks, executing actions, and using tools as part of the programming process.

OpenAI also said the new model provides clearer indications of its limitations and follows user instructions and safety requirements more closely than its predecessor.

The model is currently available to Plus, Pro, Business, Enterprise, and Edu users through ChatGPT Work and Codex.

However, OpenAI said GPT-6.1 Sol has not yet been added to the standard consumer model-selection menu across all ChatGPT interfaces.

Source: QNA